• My CV / Resume [PDF]

• News: I will go to attend NIPS2016, Barcelona, SPAIN, December 05 -- December 10, 2016

• NEWS 1: I published US patent: Speech separation method and system, US patent, US 20160189730A1, 2016

• NEWS 2: DNN based speech enhancement tool is open now and can be downloaded at https://drive.google.com/file/d/0B5r5bvRpQ5DRR1lIV1hpZ0RLQ0E/view
• https://github.com/yongxuUSTC/DNN-for-speech-enhancement

• NEWS 3: I have joined in University of Surrey, Guildford, UK as a research fellow from May, 2016.

• NEWS 4: I will go to Dresden, Germany to attend Interspeech2015 from Sept. 5, 2015 to Sept. 12, 2015. During that, i will also serve as one of the volunteers of Interspeech2015.

• NEWS 5: I will stay in Bosch Company for a week from Oct. 30, 2014, Palo Alto, California, USA. [Back to Georgia Tech on Nov. 8, 2014]

• NEWS 6: I am now at Georgia Institute of Technology for 6 months from Sep. 28, 2014, Atlanta, Georgia, USA.

• NEWS 7: One of my journal paper: "A Regression Approach to Speech Enhancement Based on Deep Neural Networks" is accepted by IEEE/ACM TRANSACTIONS ON AUDIO, SPEECH AND LANGUAGE PROCESSING.

At my master's period in USTC, i have done some deep research in speech recognition and audo index&search, especially the OOV(out of vocabulary) problem in spoken term detecion.

During my PHD in USTC (University of Science and Technology of China), my research focused on noise robust speech recognition, speech enhancement, deep neural networks (DNN), which are one of the most challenging or interesting things in speech processing.

My supervisors are Prof. Jun Du (USTC, China), Prof. Li-Rong Dai (USTC, China), and Prof. Chin-Hui Lee (Gatech, USA)

--------------------------------------------------------------------------------------

Project and source tools:

[1] Distant speech recognition

      I had successfully developed a very effective ASR acoustic model based on DNN/LSTM models for distant speech recognition to address the reverberation problem combined with microphone array based beamforming. Meanwhile i developed a reverberated speech simulation tool with the room impulse response (conv_speech tool Download here!)